"Deepa Remesh (JIRA)" <[email protected]> writes: > * Adds a test 'derbyStress.java' to jdbcapi suite. This test is > based on the repro for this patch. Without this patch, it fails when > run with client driver. Kathey had suggested in another mail that > tests for client memory leak problems (DERBY-557, DERBY-210) can be > added to same test. I did not see an existing test. So I created > this new test. If DERBY-557 does not have a test, I think it can be > added to this new test.
Hi Deepa, I think I promised to write a test for these issues once. Actually, I started writing a test for DERBY-23 (another memory leak) and I wanted to include tests for DERBY-557 and DERBY-210 in that test. But when I started I realized that the leak wasn't completely stopped. It had stopped leaking in some configurations, but in other configurations it was still leaking. Different operating systems, JVM versions, running from classes or jars - all of these factors influenced on the result (in one configuration I actually could (reproducibly) stop the memory leak by appending a dummy directory to the classpath). Finally it got so complex that I decided to put it on the shelf for a while, and it's still on the shelf... Thanks for writing this test. I will have a look at it and add a test for DERBY-557 (and possibly DERBY-23) when your patch is committed. -- Knut Anders
